1: Sample Collection and Preparation: How Do I Get My Sample Ready for GC-MS Analysis?

In order to conduct a successful GC-MS analysis there are some fundamental requirements. Gas chromatography relies on the analytes being in the gas phase; therefore, one important element is to ensure the optimal conditions are in place for this to occur. The second fundamental part is ensuring that a sufficient amount of the analyte reaches the detector. Although a mass spectrometer is a very sensitive detector, like every detector, it has limits. Therefore, any successful sample preparation must consider these limitations.

There are many different sample preparation techniques which can be used or automated for GC-MS. The best sampling or sample preparation technique to use is determined by:

• Sample phase: gas/liquid/solid or something in-between?

• Where is the sample? Can a portion be moved into the lab or must it be sampled in situ (can the instrument be taken to it)?

• Analytes: volatile/semi-volatile/involatile?

• Is it possible, if necessary, to automate the sampling/preparation technique?

A sample will be in one of three states, solid, liquid or gas. The technique used for its successful introduction into the GC-MS will vary based on the sample state. This chapter is therefore sub-divided into these three states.

Gas-phase samples are already in the state in which GC separations occur, therefore, there is no need for further transformation. Some gas-phase samples must be sampled in situ, for example air, breath from a patient or air from a processing plant. For others, a large sample can be taken and then a portion of this analysed, for example a cylinder of industrial gas, or a canister filled with the gas-phase sample.

The analytes in gas-phase samples are usually already gaseous, therefore the sampling of gases for analysis should be quite straight forward, as long as the sample is kept under leak-free conditions. However, even samples of this type have some challenges, these include:

• Enrichment of the sample to ensure the concentration is high enough to allow successful detection.

• Accurate sampling to ensure the sampled fraction is representative of the bulk.

• Transfer to the GC-MS system which does not change the sample through reaction or absorption onto surfaces and ensures the sample volume entering the GC is optimal for the gas chromatographic process, for example, delivered in a narrow band onto the head of the column.

• Storage of the sample before analysis to maintain its integrity both qualitatively and quantitatively.

• The preparation of suitable standards for quantitative analysis.

Sampling an accurate representation of the bulk is a key step. The gaseous environment can be sampled in a variety of ways and these are described in a number of standards, for example ASTM (D3588, D5466), EPA (TO-14A & TO-15, EPA Method 18) and ISO (3171).

As these standards indicate, there are a wide variety of methods used to sample a gas and it usually takes place in three different ways: spot, continuous or representative.

Spot samples are taken at one time and at one point, usually via a pitot tube. The pitot tube is a pressure measurement instrument and its primary function is to measure the fluid flow velocity of liquid, air and gas flows. It is inserted into the process (e.g. stack monitoring) or pipeline (e.g. gas outlet), or via a valve. Analysis occurs immediately without the need for storage.

There are various methods that can be employed to fill and use canisters and sampling bags, a full description of which is beyond the scope of this chapter. They all share the aim of creating a representative sample from the bulk being sampled. They all involve adding a flow controller to a sample over extended periods of time rather than immediate sampling.

For example, GPA Standard 2166 describes eight different sampling methods which are listed below:

• Evacuated container method: gas is introduced into an evacuated sample container with a pressure of less than 1 mm Hg.

• Reduced pressure method: similar to the evacuated contained method, but for higher inlet pressures.

• Helium pop method: beginning with an evacuated sample container, this is filled with helium (to around 5 psi), and then filled with the gas sample.

• Floating piston cylinder method: this method has a pre-charge chamber and sample chamber created by a piston. The pre-charge chamber is filled with an inert gas (slightly above line pressure). The outlet valve is opened and the sample displaces the piston and fills the cylinder.

• Water displacement method: the sample cylinder is filled with clean water and a vessel to measure the displaced water is attached. The gas sample is slowly introduced and the outlet valve slowly opened. The gas is sampled until all the water is displaced (detected using the sound or by observation).

• Glycol displacement method: the same as the water displacement method but using glycol rather than water.

• Purging – fill and empty method: the sample is used to purge the container, it is then emptied by releasing the output valve. This process is repeated several times to obtain a representative sample.

• Purging – controlled rate method: the rate of entry is controlled by flow controllers on the inlet and outlet.

Deactivated SUMMA™ or SilcoSteel™ canisters have been internally treated to ensure the collected analytes do not react with the stainless steel surface. SUMMA™ canisters use a passivation process to apply a nickel-chromium oxide layer, whereas SilcoSteel™ canisters have an internal silica layer. However, the term SUMMA™ canister is sometimes used to describe both types of coated canister (ASTM D5466 – 15).

When selecting a canister, the volume and nature of the sample, and the location that the sample is being taken needs to be taken into consideration.

Before sampling, unused bags should be stored in a clean environment and sealed in an outer bag to prevent adsorption of contaminants. Bags should be pre-cleaned before use by flushing with high-purity nitrogen. For validation, compounds must be stable in the bag or canister over the period in which the validation is conducted. Overall, the leak rate from the bag must be low.

During sampling it should be ensured that any tubing used for the bag connections is clean. A known and predictable flow rate should be used. Bags should not be overfilled, no more than 80% of the stated maximum bag volume should be filled.

Bags are intended for a single use, owing to potential sample adsorption onto the bag film. Therefore, after sampling and analysis, best practice is to not re-use bags. Hold times in bags before analysis are typically recommended to be 48 h or less owing to concerns around adsorption onto bag surfaces, unless the validation study demonstrates a longer stability. Bags containing samples should be protected from direct sunlight and stored above 0 °C to prevent condensation. Bags should be transported in rigid containers to prevent bag puncture and not shipped by air unless samples will be kept in a pressurised area.

Gas-phase samples can be actively sampled in situ by drawing the sample through a conditioned sorbent and packed into a thermal desorption (TD) tube using a constant pressure or constant flow pump. The tube is then sealed, to prevent the loss of analytes and the ingress of contaminants, then returned to the lab for analysis. Tubes can be stable for several weeks. For example, parts-per-trillion (ppt) levels of polyaromatic hydrocarbons (PAHs) can be detected using air analysis, by drawing 100 L of air through a packed TD tube. See Section 1.3.1 for background information on thermal desorption.

The solid sorbent can also be held within a device such as an ORBO™ tube or filter. An impinger enables collection with a liquid sorbent.

Thermal desorption (TD) is a physical separation process, in which heat is applied to a sample to transfer analytes that are adsorbed or absorbed within the sample tube, into the gas-phase so that they can be analysed using gas chromatography. TD only uses temperatures up to 350 °C and therefore no chemical bonds are broken in the process, only interactions. TD can be used to analyse a range of species, from those as volatile as acetylene (with two carbon atoms) up to molecules with forty carbons, such as PAHs and phthalates.

Small solid or viscous liquid samples can be directly thermally desorbed by placing them in a conditioned TD tube (see Section 1.3.1). As mentioned previously, gas-phase analytes can be concentrated by drawing the gas-phase sample through a conditioned TD tube packed with a sorbent.

The TD tube itself can be made from: glass, which is beneficial for observation of the position of solid samples placed directly into the tube (see Section 1.3.1); stainless steel, which makes it very robust, especially for those tubes sampled away from the lab; or coated (silco) steel, which makes the tube very inert and is much better for active analytes such as those molecules containing sulphur.

The tubes vary in size depending on the manufacturer, but the industry standard TD methods, such as ISO, CEN, ASTM and EPA, use a 3.5 × ¼ in. outside diameter (o.d.) tubes. Tubes should have a unique identifier, which enables the sample to be matched to the tube and the sampling direction must be known so that the tube is desorbed in the reverse direction to ensure that all the sampled analytes are recovered.

Similar to solid phase micro-extraction (SPME), the sorbents placed into the TD tube and the cold trap can either interact with the analytes through absorption or adsorption. The sorbent(s) selected is dependent on the target analytes. It must trap the target analytes at the ambient temperature of the sampling location and easily release them again when rapidly heated. This temperature must not be higher than the maximum temperature of the sorbent, with no irreversible ad/absorption or catalytic breakdown.

Common sorbents are polymers such as Tenax ®, Porapak, Hayesep or Chromosorb, a styrene divinylbenzene (DVB) polymer; carbon molecular sieves such as Sulficarb, Carbosieve or Carboxen; zeolite molecular sieves; or graphitised carbon black such as Carbopack, Carbotrap or Carbograph. Tenax® and graphitised carbon blacks are hydrophobic and are therefore beneficial for ‘wet’ samples. Carbon molecular sieves are mostly hydrophilic with Carboxen being the most hydrophobic. Zeolite molecular sieves are very hydrophilic and can collect water up to the mg level, in a typically sized TD tube.

Different types of sorbents are good for different volatilities and polarities of analytes and have different retention volumes and maximum temperatures (ranging from 190–400 °C). Even when the maximum temperatures are not exceeded, some sorbents can produce artefacts, that is the release of molecules, that are focused, separated and detected using GC. Carbon molecular sieves have minimal artefact levels, whereas Tenax® has a low artefact level when new, but the artefact level increases as the sorbent ages.

Porous polymers and carbonised molecular sieves are more inert than graphitised carbon blacks such as Carbograph 1TD. Generally, the more volatile the analyte(s), the stronger the sorbent must be. For analytes with a boiling point (bp) greater than 100 °C a weak sorbent such as Tenax® TA is used. Those analytes with a boiling point between 30–100 °C require the use of a medium strength sorbent such as Carbograph 1TD. Very volatile analytes with a boiling point between 30–50 °C require the of use a strong sorbent such as Sulficarb or Carboxen 1000.

The mesh size of the packing material affects the packing density and the back pressure that it creates. A mesh size of 20/40 has larger particles than 60/80 and therefore can be sampled using higher flow rates.

The sorbent life of the tube is dependent on the type of sorbent(s) used, the maximum and routine desorption temperatures that it has been exposed to and the number of desorption cycles, which includes conditioning of the tubes. Tenax® and carbon-based sorbents are usually good for 100–200 cycles, whereas porous polymers are less stable, usually with a lifetime of 100 cycles.

A trap is used to collect and focus analytes between the TD tube and the GC analytical column. Without this, the long TD tube desorption times would result in long transfer times to the column, producing broad sample bands and poor chromatographic resolution and peak shapes. The analytes can be trapped through cryofocusing in the inlet liner or on a GC pre-column using cryogens or through cold trapping with a Peltier cooled trap (cold trap). The cold trap enables the use of a small amount of sorbent in a narrow tube to selectively concentrate the analytes, but unlike the TD tube, the cold trap is cooled below the ambient temperature, reducing the likelihood of break-through even though less sorbent is used. With cryofocusing, everything released from the TD tube is trapped, but with the cold trap the sorbent can be selected so as not to trap unwanted gases such as water, solvents, and so forth. As with the TD tube sorbent, it must trap the target analytes at the (lower) temperature chosen and then easily release them with no catalytic breakdown when the trap is rapidly heated. As the cold trap is usually backflushed, multiple sorbents can be chosen, sometimes up to three, to match the target analytes. Upon rapidly heating the narrow cold trap, the analytes are typically transferred to the GC column in split mode to increase the flow through the cold trap when desorbing, resulting in a faster transfer to the GC column and therefore producing a narrow sample band. Even with a low split ratio, the sample flow to waste is higher than that onto the column resulting in a fraction of the sample being separated on the GC. Recently developed instrumentation can automatically enable the split flow effluent from the cold trap desorption to pass back through the original sample TD tube or through a new, conditioned TD tube to re-collect the sample. This means that TD samples are no longer one-shot, where if something went wrong with the analysis of the sample it could not be re-analysed. In addition, the recovery is quantitative, therefore if the sample does have to be re-analysed the original concentration can be determined based on the split ratio. Sample re-collection can also be achieved through the trapping of the split effluent from the TD tube desorption onto a new, conditioned TD tube if a split method was used in that step.

After placing the sample in the tube or, concentrating the gas-phase sample in a packed tube, the tube is placed in the TD instrument and checked for leaks. Upon passing, it is pre-purged with a carrier gas, usually helium, to remove atmospheric oxygen and to prevent oxidation of the sample upon heating. The TD tube is then heated to a temperature dependent on the maximum temperature of the sorbent, the volatility of the analytes and the nature of the sample. It is then held at that temperature, usually for 5–30 minutes, to fully desorb the sample or sorbent. The TD tube is continuously purged while being heated and the emerging analytes are backflushed off the TD tube and selectively concentrated on a narrow cold (focusing) trap, containing a small amount of sorbent. During this process, unwanted gases, water and matrix are removed. If the sample is very concentrated, the analytes can be transferred to the cold trap in the split mode, rather than the usual splitless mode. After the TD tube is fully desorbed, the cold trap is rapidly heated and the concentrated analytes are backflushed through a heated transfer line, usually as a split injection, onto the GC analytical column in a narrow sample band for separation and detection.

Before use, TD tubes must be conditioned and sealed before being taken to the sampling location. Those packed with a sorbent must be conditioned at the optimal temperature for the required time for the specific sorbent(s) used.

The break-through volume for the samples and analytes to be sampled must be determined before sample collection, so that a known volume of the sampled gas can be taken which is below the break-through volume. The break-through volume can be simply checked by attaching two sampling tubes in series; if the first tube becomes saturated any analytes breaking-through will be trapped by the second tube, this can then be analysed to determine if the break-through volume has been exceeded.

During sampling, the optimal flow rate through the tube affects the interaction of the analyte with the sorbent and therefore the amount that can be trapped. The optimal sampling rate through a standard 5 mm inside diameter (i.d.) packed TD tube is 50 mL min⁻¹, with a working range of 10–200 mL min⁻¹ that can increase to 500 mL min⁻¹ for a maximum of 10 to 15 min. Sample volumes range from 500 mL to 100 L.

Improved adsorption efficiencies are observed at lower temperatures, however the ambient temperature during sampling must be considered when selecting the sorbent(s). If there is a wide range of physical and chemical properties of the analytes, a tube can be packed with multiple sorbent beds, usually up to a maximum of three, so that there is a sufficient mass of each sorbent for the capacity required. The weakest sorbent is packed closest to the sample, with the strongest sorbent packed furthest from the sampling end of the tube to enable the fast and efficient release of the analytes in the reverse direction. The analytes are then back-flushed to the next stage of the TD process.

If replicate sample aliquots are required, a manifold sampling system can be used in which there are multiple tubes, either containing the same sorbent(s) for repeatability checks or containing different sorbents to sample different analytes. These are set-up in parallel and are all connected to the same pump, but with flow control valves on each, to either balance the flows or to set individual flows for each TD tube.

Quantitation can be performed, as with other techniques (see Chapter 8), by analysing the standards using the same method as the samples. A maximum of 1–2 µL of standard solution, preferably in a solvent that is not trapped by the sorbent and has a low vapour expansion coefficient, or its headspace is directly spiked onto the TD tube using a spiking rig. The rig enables the standard to be injected into a carrier gas so that the analytes are blown through the TD tube sorbent to be trapped in the same manner as the sample analytes and then any solvent is purged before analysis. The automatic addition of an internal standard (IS) into the TD tube and/or the cold trap is possible and is instrument dependent.

Passive sampling is achieved when contaminated air enters the device and diffuses through the sorbent without the use of any pumps. The sorbent can then be analysed by thermally desorbing the analytes into a GC-MS system, for example via a TD instrument or with liquid desorption and injection.

Passive sampling is frequently used for occupational health to monitor indoor and ambient air. A packed TD tube with an axial diffusive sampler is used for long-term exposure, from 8 hours to 4 weeks. Radial diffusive samplers have a much larger surface area and therefore are good for short-term monitoring of 0.5 to 6 hours. After sampling, the cartridge is removed from the sampler and placed in an empty TD tube for analysis. Monitoring badges are similar to radial diffusive samplers. Standard methods present compound-specific diffusion uptake rate data for different devices, which is used to quantify the concentration of analytes in the air.

Thermal desorption instruments can also be configured for online sampling, in which the gas-phase analytes are trapped, concentrated and analysed. The packed TD tubes themselves are not needed, as the sample is pumped directly through the cold (focusing) trap. Two traps in parallel enable continuous monitoring and they can operate, unattended, for days, weeks or even years! These are commercially available from a number of suppliers including Markes International or Perkin-Elmer.

Some liquid-phase samples may need to be sampled in situ, for example process and waste streams. For others, a sub-sample may be taken and then a portion of this analysed in the lab, for example water, chemicals, urine or fuel. The analytes in liquid-phase samples can range from volatile to involatile and therefore care must be taken not to lose these when sampling, transporting and storing.

As with a gaseous sample the same key points need to be considered:

• Storage of the sample before analysis to maintain its integrity both qualitatively and quantitatively.

• Dilution or enrichment of the sample analytes to ensure concentration is high enough to allow successful detection, the concentration should be in the working range of the GC-MS method. Lower or higher analyte concentrations can be obtained through several stages of the analysis process:

  • ○ During sample preparation through dilution with a solvent or by using a sample enrichment technique or solvent blow-down concentration step under nitrogen.

  • ○ By injecting a very small (e.g. 0.1 µL) or large volume (up to 1000 µL) of a liquid sample, see Chapter 2.

  • ○ During sample introduction to the column by performing split or splitless injection, see Chapter 2.

• Transfer the sample into the GC system in a way that does not change the sample through reaction or adsorption onto surfaces and ensures the sample entering the GC is optimal for the gas chromatography process, that is, it is delivered as a narrow band onto the head of the column, see Chapter 2.

Successful storage of a liquid sample is very application dependent, however, broadly speaking the integrity of samples can be maintained by slowing or eliminating sources of chemical or biological degradation. This may include:

• Protection from light, to prevent photo-degradation.

• Protection from high temperatures to minimise evaporation of volatile analytes, thermal degradation and growth of microbes.

• Protection from the ingress of reactive components such as oxygen or moisture.

The extent of protection required is established on a case-by-case basis. However, for general protection a well-sealed container is used, purged if necessary with dry nitrogen to remove reactive oxygen and stored in a fridge, cupboard or other monitored environment.

Whether diluting a liquid sample for direct injection, or using a solvent in the sample preparation process a solvent should be selected that is appropriate for both the sample and the analysis method:

• Choose a solvent which dissolves the analytes of interest but does not react with them. This probably means a solvent which is similar in nature to the sample. It should also be miscible with the sample solvent, otherwise this will result in a liquid–liquid extraction (LLE).

• Choose a solvent that is GC-amenable that:

  • ○ Enables the GC to deliver sharp, well resolved peaks in which the solvent is either eluted quickly at the start of the chromatogram or eluted later after the peaks of interest are separated and eluted. The former is more common and enables easier method development than the latter arrangement, although for some methods the solvent elutes between the peaks but this should be avoided if possible.

  • ○ Is suitable for the sample introduction technique being used. This includes having a boiling point (b.p.) that is relatively lower than the most volatile analytes to be determined. The lowest initial oven temperature that is attainable without cryogenic cooling in the environment in which the GC is sited should also be considered if solvent re-condensing on the column is required. For example, if dichloromethane (b.p. 39 °C) is used for an on-column, splitless or large volume injection, a low oven initial temperature at a maximum of 35 °C, preferably 30 °C, must be easily attainable. Another factor is the solvent polarity relative to the inlet liner packing material and column stationary phase, see Chapter 2 for more information on these points.

  • ○ It should not react with any GC components, for example the injection syringe, inlet septum or liner, column stationary phase or detector. This can lead to leaks where it reacts with seals such as the septum, contamination producing additional peaks in the chromatogram, high bleed which produces high baselines, and can ultimately result in higher consumable costs, poor chromatography and sensitivity.

Therefore, typically the solvent is volatile with a lower boiling point than the most volatile analyte. This allows the widest range of options when setting parameters for the GC and also allows the start of the MS data collection to be delayed until after this solvent peak has cleared the vacuum system, this is called the solvent delay. The MS filaments within the ion source are turned off until after the solvent delay, to prevent them from being overloaded by this large quantity of solvent which will cause them to prematurely fail. The pressure in the mass spectrometer is also higher at this point which can affect the performance of the mass spectrometer if it is acquiring a spectrum, thereby reducing sensitivity. Information on how to optimise the solvent delay time is discussed in Chapter 5.

Very few liquid (or solid) samples can be directly introduced into a GC without some form of preparation. The reasons for sample preparation depend on the nature of the sample itself and can include one or all of the following examples:

• Removal of interfering matrix components which could prevent the detection of the analytes.

• Protection of the analytical system from contaminants, which would cause frequent maintenance and high consumable costs.

• Selective isolation and concentration of the analytes with 100–5000x enrichment.

• Changing the phase or the sample solvent: for example, water samples are very difficult to directly inject.

The transition from the liquid to the gas phase places some limitations on the volume of sample which can be introduced into the GC. It is therefore important to consider if sample enrichment will be required in order to ensure that the amount of sample reaching the detector is sufficient for detection. The typical amount which can be detected using MS will vary and is dependent on the type and design of the mass spectrometer; this is discussed in later chapters. However, a typical single quadrupole MS operating in full scan mode requires around 1–100 pg to reach the detector from the column. If 1 µL is introduced into the GC then the concentration of this solution needs to be at least 1–100 pg per µL, which is equal to 1–100 ng mL⁻¹ or 1–100 ppb. This assumes that the MS is working correctly, with regular checks being performed for MS tuning and calibration of the method (details of which are covered in later chapters).

Sample preparation can be performed:

• Manually, before the liquid extract is placed into an autosampler vial for manual injection or via a standard liquid autosampler for injection into the GC.

• Automatically with a special autosampler:

  • ○ For example, an XYZ robot can be programmed to perform each of the sample preparation steps which are usually performed manually.

  • ○ Or an autosampler designed to only perform the specific sampling or sample preparation technique, for example headspace analysis.

The main methods used to prepare a liquid sample for GC-MS analysis are:

• Using a liquid phase.

• Using a solid phase.

• Thermally, by heating the sample.

Liquid–liquid extraction is the mass-transfer of analyte(s) from one liquid phase into another liquid phase. The concentration of the analyte is based on the affinity of the analyte for the two immiscible liquids, usually water and an organic solvent.

The volatile and semi-volatile analytes present in a solution or a liquid sample can be enriched by extraction into another liquid. Enrichment can occur as it is possible to extract a large volume of the solution or liquid sample into a much lower volume of a second solvent. This can have the added benefit of removing any unwanted matrix (e.g. salts or proteins). This process of liquid–liquid extraction (LLE) has traditionally been conducted in a glass separating funnel. However, variations on this basic technique have been used in a wide variety of ways to optimise the enrichment ratio, including automating the entire process to GC-MS with the use of XYZ robots.

In traditional LLE the original solution (typically water) is extracted with an immiscible solvent (for example dichloromethane). The analytes of interest (typically non-polar in nature) have a greater solubility in the organic phase than the aqueous phase therefore partitioning of the analytes occurs whereby the majority of them move from the aqueous phase into the organic solvent. This process (when conducted in a separating funnel) is facilitated by shaking of the funnel to encourage mixing of the phases and movement of the analytes between the phases. The phases are then allowed to separate and the organic phase is decanted from the aqueous phase. The separating funnel facilitates this removal by a tap in the base of the funnel through which controlled quantities of the lower phase can be removed. The process is normally repeated at least three times with fresh organic solvent being used to partition the analyte completely into the organic portions.

This basic LLE process has been explored in many formats. Many of which aim to increase the enrichment process, make the partition process quicker or improve the extraction efficiency for analytes which have only a partial affinity for the organic/collection phase. For example, matrix modification is used to improve extraction and includes salting out and changing the pH, these are discussed further in this chapter under headspace analysis. This enrichment process has also been further refined and adapted to broaden its application and make use of smaller sample volumes to enable automation. A few of the more common adaptations are discussed below.

In this adaption of the basic LLE, the liquid sample (typically water) has the extraction medium, typically a more dense solvent such as chloroform or dichloromethane (extraction solvent), rapidly introduced into the sample with a third solvent (the disperser) to aid the rapid and complete dispersion of the extraction solvent throughout the bulk of the sample solution. This is normally achieved by the use of a microliter syringe to introduce the solvents rapidly (less than 1 sec). The highly dispersed extraction solvent allows rapid partitioning of the analytes to occur between the liquid phases with potentially a higher efficiency than classical LLE owing to the very large surface area between the sample solvent and the extraction solvent. When the partitioning is complete, the extraction solvent is recovered by centrifugation of the sample separating the liquid phases. The extraction solvent phase is then removed and injected into the GC-MS for analysis (Figure 1.1). For this technique to be successful the precise ratio of the sample solvent, dispenser solvent and extraction solvent must be optimised.

In this technique, an immiscible solvent droplet is held either in the sealed heated headspace above the sample or lowered into the bulk of the liquid sample from a microlitre syringe and is held suspended in this position whilst analytes from the sample partition into the solvent droplet. As with DLLME, the droplet presents a larger surface area to the sample and thus makes the liquid–liquid (or gas liquid partitioning in the case of headspace sampling) quicker than traditional LLE. It also operates at a much smaller scale. This approach uses microlitres of extraction solvent and 1–10 mL of sample. This means that the enrichment factors are high. That is, trace quantities of analytes in the sample are transferred into a few microlitres and injected onto the GC. Rather than these same quantities being transferred into a few millilitres of solvent then only a few microlitres being injected into the GC, with the bulk of the analytes not used in the analysis. There are a number of variables to control and optimise in this process such as the choice of extraction solvent, temperature of extraction and the extraction time.

Solid phase extraction is a separation method that uses a solid phase to extract analyte(s) from a sample, based on their preferential affinity for the solid material over the sample.

Solid Phase Extraction (SPE) and its related version dispersive SPE (dSPE), have long been used as extraction, clean-up and concentration processes and can perform all preparation tasks simultaneously. It uses solid sorbents to selectively remove either target analytes or unwanted interferences from samples prior to analysis. For example, contaminants can be eliminated through the following mechanisms:

• They are unretained and pass through during the loading step.

• Contaminants that are retained but are more polar than the analytes can be washed off.

• Contaminants that are less polar than the analytes can be left bound to the sorbent, for example, this is a good way of eliminating lipids from biological samples.

There are a number of steps involved in sample preparation using SPE, the typical principles include:

1. Sample pre-treatment: can include filtration to remove particles, changing pH to remove ionic interferences and dilution to minimise ionic and non-polar interferences.

2. Conditioning: clean the sorbent with 3–5 sorbent bed volumes of the analyte elution solvent.

3. Equilibrating: prepare the sorbent with 3–5 bed volumes of the sample solvent (known as solvation or wetting). Critical for a good efficiency and to obtaining the optimal sorbent capacity, otherwise this can result in break-through and poor recovery.

4. Loading: the sample is pushed or pulled through the sorbent, extracting and enriching analytes and some matrix components. A slow flow rate is often better, as too rapid a rate can lead to poor retention and recovery of analytes.

5. Air drying: can be used to remove residual water (optional).

6. Washing: analytes are retained and remaining interferents washed out. The solvent strength and pH must be optimised for selectivity, two half volume wash steps often work better than one entire volume.

7. Elution: analytes are eluted with a selective solvent and remaining interferents are retained. The solvent must be optimised and should be a stronger solvent than that used in the wash step and the elution volume should be determined. Large volumes can be evaporated (another step to optimise) ensuring volatile analytes are not lost, or a large volume injection into the GC may be performed.

Particulates will not pass through the SPE sorbent. Low levels will sit on top of the sorbent and can be extracted, but high levels should be pre-filtered before SPE if they make up a high percentage of the sample, otherwise they will block the cartridge.

There are a variety of SPE sorbent holders:

• Cartridges are used for the elution of small sample volumes (1–60 mL) with an optimal flow of 1–10 mL min⁻¹. They have a small surface area but a long path length and are suitable for low level particulates. There are multiple formats:

  • ○ straight barrel;

  • ○ large reservoir capacity;

  • ○ Luer cartridges that can be run in series;

  • ○ 96 well plate format;

  • ○ micro-syringe format with an integrated SPE cartridge in the syringe needle (MEPS™).

• Columns are similar to cartridges, but can come in very large formats for preparative applications.

• Disks have a large surface area (47, 50 or 90 mm in diameter) and have faster optimal flows through the sorbent (200 mL min⁻¹) which are better for large volumes of sample (more than 100 mL) or samples with a higher proportion of particulates to be extracted.

There are a wide range of solid sorbents to choose from and the sample may be extracted through multiple sorbent beds, for example to selectively extract polar analytes and then non-polar analytes. The two main classes of sorbents are:

• Silica-based:

  • ○ Non-polar C18, C8 and Ph modified silicas: used for non-polar analytes, with retention through dispersion or hydrophobic/hydrophilic interactions; these analytes are easily adsorbed in a polar environment; they are eluted using a solvent of lower polarity.

  • ○ Polar unmodified silica (-OH groups), CN, NH₂ and –OH (diol): used for polar analytes with amino, hydroxyl, carbonyl, hetero-atoms, aromatic rings, double bonds and so forth, with retention through hydrogen bonding, dipole–dipole and π–π interactions; these analytes are easily adsorbed in a non-polar environment and they are eluted using a polar solvent.

  • ○ Ionic:

    • ■ Cation exchange: silica modified with an acid (e.g. benzene sulphonic acid). Retention is through the interaction of charged, cationic groups on the analytes of interest and charged, anionic (negative) functional groups on the sorbent, via electrostatic (ionic) interactions. They are eluted with a solvent of high ionic strength, in which the presence of a large number of ions in the elution buffer creates competition with the analyte groups for the sorbent groups or through a pH change, in which either the charged analyte groups are neutralised, or the charged sorbent groups are neutralised, or through use of a buffer containing cationic species with a high affinity for the sorbent functional groups.

    • ■ Anion exchange: silica modified with a quaternary amine, used for analytes with cationic groups: 1°, 2°, 3° and 4° amines and inorganic cations, and analytes with anionic groups: carboxylic and sulphonic acids, phosphates and so forth. The sorbent extracts analytes based on electrostatic interactions between the analyte of interest and the positively charged groups on the stationary phase. For ion exchange to occur, both the stationary phase and the sample must be at a pH in which both are charged. Ion exchange interactions are enhanced in low ionic strength samples and with counter ions with low selectivity (e.g. acetate, Na⁺), they are eluted using a solvent of high ionic strength and high selectivity (e.g. citrate or Ca²⁺).

    • ■ Mixed mode columns: contain both ion exchange and hydrophobic ligands on a sorbent surface.

• Polymer-based:

  • ○ Polymer or resin type: typically co-polymers with a single hydrophobic monomer plus a more polar monomer, can include a weak/strong ion exchange functionality, for example divinyl benzene resin with sulphonate groups. They have a higher loading (10–20% w/w) than silica-based sorbents (1–5% w/w).

  • ○ Molecularly imprinted polymers (MIPs): highly cross-linked polymer-based molecular recognition elements that are engineered to bind one target or a class of structurally related analytes. During synthesis a template molecule guides the formation of specific cavities in the polymer which are sterically and chemically complementary to the analytes with multiple non-covalent interactions. They are highly selective and use harsh wash conditions during sample preparation to obtain a lower background and therefore lower detection limits.

To develop an SPE method, there are a number of steps that must be followed:

1. Determine the structure and solubility of the analytes: it is necessary to determine the SPE sorbent to be used as this can help with solvent selection. Non-polar analytes have a Log P of more than 4.0; mid-polar analytes have a Log P of 1.5–4.0; and polar analytes have a Log P of less than 1.5. In reversed phase SPE (RP-SPE), the Log P value of the analyte must be greater than 1.5 in order to be retained on the non-polar sorbent. For ionisable analytes it is best to neutralise their charge prior to loading.

2. Select an appropriate SPE sorbent and weight: the capacity of the sorbent is the total amount of analyte that can be adsorbed under optimal conditions. This is 3–5% of the sorbent weight for silica sorbents and up to 30% for some polymers. Capacities should not be exceeded, otherwise the method will not be quantitative.

3. Determine the sample concentration and maximum loading.

4. Determine the optimal wash solvent: wash with as strong a solvent as possible that will not elute the analytes.

5. Determine the optimal elution solvent: should be optimised first using standards and no matrix interference. For example, for RP-SPE make up a range of elution solvents with 0–100% of organic solvents in 10% steps and use these to elute each cartridge analysing and plotting the results.

   • At 100% organic solvents the recovery should be 80–100%, if not a stronger elution solvent is required or a different sorbent.

   • The lowest percentage of organic solvent with approximately 0% recovery is the best wash solvent.

   • The lowest percentage of organic solvent with the highest recovery is the best elution solvent, leaving the less polar contaminants retained on the sorbent.

In some ways this technique is similar to SDME, however the liquid extraction medium is a high molecular weight liquid or a solid sorbent coated onto a fibre, both of which are similar to a GC column stationary phase. This is a solventless extraction technique. As with SDME, the fibre is exposed either to the headspace above the sample (headspace-SPME) or the liquid bulk (direct immersion-SPME) to extract and concentrate the analytes, as illustrated in Figure 1.2. Analytes from the liquid are partitioned or adsorbed from the sample onto the coated fibre. Once extraction is completed, the fibre is removed from the sample and either introduced into the injection port of the GC where the analytes are thermally desorbed or liquid desorption takes place with a solvent.

The fibre is formed of 1–2 cm of fused silica coated with a stationary phase and bonded to a stainless-steel plunger. When piercing the vial or the GC inlet septum, the plunger is withdrawn within the needle to protect the fibre and then exposed to extract analytes or desorb the fibre. The phase type and thickness is chosen to match the characteristics of the analytes (selectivity). The amount of analyte adsorbed depends on the thickness of the phase and the partition coefficient of the analyte. As for GC columns, a thick phase is chosen for volatile analytes and a thin phase for semi-volatile analytes. There are two broad types of classification of SPME fibre phases:

• Homogeneous pure polymer coatings produce absorptive fibres in which the extraction occurs through partitioning, resulting in a good capacity and broad linearity:

  • ○ Polydimethylsiloxane (PDMS):

    • ■ 100 µm volatiles (MW 60–275)

    • ■ 30 µm non-polar semi-volatiles (MW 80–500)

    • ■ 7 µm mid- to non-polar semi-volatiles (MW 125–600)

  • ○ Carbowax (PEG): 60 µm for alcohols & polar volatiles (MW 40–275)

  • ○ Polyacrylate (PA): 85 µm for polar semi-volatiles (MW 80–300)

• Porous particles imbedded in a partially crosslinked polymeric phase producing both absorptive and adsorptive fibres which are better for extracting low concentration analytes:

  • ○ PDMS/DVB: 65 µm for polar volatiles, amines and nitro-aromatics (MW 50–300)

  • ○ Carboxen (CAR)/PDMS: 75/85 µm for trace-level volatiles (MW 30–225)

  • ○ DVB/CAR/PDMS: 50/30 µm for flavour compounds C3–C20 (MW 40–275)

  • ○ Activated carbon fibre (ACF): varying polarities, high temperature phase (380 °C).

The addition of a sorbent to a coating, for example strongly polar Carbowax PEG onto DVB polymer, increases the surface area and improves the extraction efficiency of polar molecules.

SPME analysis can be classified into two types, depending on where the fibre extracts the sample from:

• Headspace SPME (HS-SPME) extracts the analytes from the headspace above the sample and is therefore more sensitive for volatile analytes. The analytes must be predominantly in the headspace to be extracted, therefore matrix modification may be required to release the analytes into the headspace, this is discussed further later in this chapter concerning headspace analysis. HS-SPME equilibrates faster and keeps the fibre cleaner, minimising interferents and prolongs fibre life.

• Direct immersion SPME (DI-SPME) inserts the fibre into the sample and therefore is more sensitive for analytes predominantly in the liquid sample. The sample must not damage the SPME fibre when extracting the analytes, as it is delicate and easily breakable.

Optimisation is again required to ensure that extraction from the sample into the fibre occurs in a reproducible manner. The method development steps are:

1. Select the most suitable SPME phases based on the chemistry of the analytes and phases:

   • Polarity: like separates like, therefore select a non-polar fibre for non-polar analytes and a polar fibre for polar analytes. Consider hydrogen-bonding, dipole–dipole, induced dipole, π–π and dispersion (non-polar) interactions.

   • Molecular weight: for example when selecting GC column stationary phases, choose a thicker phase for more volatile analytes (e.g. 100 µm) and a thinner phase for higher MW analytes (e.g. 7 µm).

2. Fibre conditioning: for the recommended temperature and time for the fibre phase chosen.

3. Sample preparation: for example, matrix modification for HS-SPME or sample dilution for DI-SPME for viscous samples.

4. Sample pre-incubation: heating and shaking temperature and time. In HS-SPME to establish an equilibrium between the sample and the headspace, the temperature must be optimised below the sample solvent boiling point. DI-SPME uses lower temperatures to minimise analytes in the headspace, but temperatures above room temperature to ensure reproducibility.

5. Extraction: the time the fibre is placed into the headspace or sample to extract the analytes, usually at the same temperature as the sample pre-incubation but with a gentle shaking at 100 rpm. The sample vial septum thickness should be rated for SPME, otherwise septum coring and fibre needle blockage may occur. The extraction time is generally longer for DI-SPME.

6. Desorption: the fibre is exposed in the GC inlet to quickly desorb analytes onto the column in 1–3 min. A narrow inlet liner with an internal diameter of 0.75 mm must be used in the GC inlet to obtain sharp peak shapes. Use of a pre-drilled inlet septum is recommended to prevent fibre breakage. The GC inlet temperature is optimised to desorb the least volatile analytes but must be no higher than the recommended maximum operating temperature for the fibre, otherwise damage will occur.

7. Fibre bake-out: the fibre must be re-conditioned post-analysis to clean it for the next sample, it is recommended to bake-out the fibre at the fibre conditioning temperature, if a fibre-conditioning station is available.

There is a linear relationship between the initial concentration of the analyte in the sample and the amount ad/absorbed, therefore the technique is quantitative, however, it is highly recommended to use an internal standard (IS) to improve the accuracy. The amount of analyte extracted is not related to the sample volume, therefore the technique can be used for field sampling of lakes, air, and so forth. However, for high concentration samples the sample volume should be no greater than 5 mL if calibrating, as the amount of analyte removed by the fibre is not sufficient to change the sample concentration.

The QuEChERS method has gained popularity for its wide applicability in the preparation of samples being analysed for pesticide residues in food. It has been adopted in a number of standards for analysing pesticide residues including EN15662 and the AOAC Official method 2007.01 as an alternative to more traditional LLE. Its key advantage is a reduction in solvent use and ease of automation, which reduces costs overall. It is a standardised approach, made up of three steps:

1. Sample preparation: the sample is ground or homogenised and an IS is added.

2. Sample extract clean-up: sample extraction and clean-up vary somewhat depending on the sample and target analytes however, the methodology is defined by the addition of an extraction solvent, buffering system (such as sodium acetate or acetic acid) and drying agent (typically magnesium sulphate). The organic layer is then cleaned-up further by dSPE using a primary secondary amine (PSA) or a C18 (octadecyl) modified silica-based or graphitized carbon black (GCB) sorbent.

3. Sample analysis: centrifugation is used to separate the solids from the liquid extracts, which are then injected into a GC-MS.

Heat can be applied directly to a liquid-phase sample to release the volatile analytes for analysis. The main techniques used for liquid samples, in order of increasing sensitivity, are static headspace, dynamic headspace and purge-and-trap analysis.

Headspace (HS) can be defined as the gas space above a sample when it is placed in a chromatography vial and therefore it is the analysis of the analytes present within that gas. Samples that can be analysed using HS analysis include anything that can fit in the vial and release volatile compounds, these include liquids and solids; here we will focus on liquid samples. Usually, heat is applied to the vial to assist the release of the compounds. How much heat can be applied depends on the boiling point of any solvents in the sample and the maximum temperature of the instrument and the vial. For most HS analyses this is usually in the thermal desorption region, below 350 °C, in which carbon–carbon bonds are not broken.

The sample is placed into a HS screw-top or crimp-top vial, ranging in size from 2–22 mL depending on the HS autosampler, with 10 or 20 mL vials being typically used. The size is important when considering the phase ratio, as detailed below. The HS vial containing the sample is then sealed with a septum and cap. The septum is usually silicone-PTFE, but can also be rubber, it should be selected to ensure that it:

• Provides a gas-tight seal;

• Does not absorb the analytes from the sample;

• Does not add bleed compounds to the vial HS.

PTFE (polytetrafluoroethylene) is very inert and should be facing the sample, however once pierced there is a route to the silicone which can bleed or absorb analytes. If possible, it is best to not pierce the septum when adding to the vial, for example when adding matrix modification reagents or spiking with an IS, these should be added before capping. Using the wrong septum for an application can have a large effect on the reproducibility and sensitivity. The vial, cap and septum should all be rated to the temperatures used in the method, especially if reactions, such as derivatisation, will take place in the vial. In which case, it is recommended to use a ‘venting’ septum, which has a weakness in case of too much vial pressure as it is preferable for the septum to break rather than the vial!

Once the sample, any standards and matrix modification reagents are sealed in the HS vial, it is then usually heated and shaken. The heating of the system causes the equilibrium of volatile analytes within the liquid to shift towards the gas phase in accordance with their relative affinity for the liquid and gas phases, shaking reduces the time taken to achieve this. The equilibrium is dependent on a number of factors which are beyond the scope of this chapter, but are referenced at the end under further reading. In simple terms, the most volatile analytes will partition towards the gas phase and the least volatile will remain in the liquid phase. The concentration of the analytes in the headspace will be affected by a number of factors including the pressure, ionic strength (in the liquid phase) and temperature, plus the length of time required for an equilibrium to be established. Matrix modification, for example the addition of salt, a change of pH, addition of a co-solvent or derivatisation can help particular analytes to move into the gas phase, for example salt is useful for enabling polar analytes to migrate out of a polar liquid matrix. This partition process is a dynamic equilibrium. At a fixed temperature, the partition coefficient is defined as the ratio of a given analyte in the gas and liquid phases. The ratio between the volume of the sample and volume of HS (phase ratio) has an impact on the recovery of those analytes with good partition coefficients, with larger sample volumes and smaller HS volumes giving this technique better sensitivity. Analytically, the aim is to migrate the volatile analytes of interest as much as possible into the gas phase whilst leaving behind the potentially interfering non-volatile sample constituents. To do this the equilibration time, temperature, matrix modification and phase ratio all need to be optimised to produce a sensitive, reproducible and robust method, the use of an IS is always recommended.

The HS can then be sampled using a gas tight syringe or heated valve and transfer line to transfer the analytes out of the HS vial and into the GC for analysis. Sampling of the HS upsets the equilibrium and each sample taken from the HS vial will result in an overall reduction in the analyte concentration.

A variation on this theme is the concept of dynamic headspace (DHS) analysis. DHS takes place in a purged vial. A flow of carrier gas is purged over the sample and continuously transfers the volatile analytes emerging from the sample into a trap, in which they are concentrated. The equilibrium is never reached and after a fixed period the purging stops and the trap is thermally desorbed, to transfer the analytes into the GC column for analysis. Alternatively, it can be eluted with an organic solvent for analysis using different techniques. As most of the analyte molecules are recovered, DHS is more sensitive than static HS, but only for those analytes with low partition coefficients. The sample is treated in the same way as for static HS, with heating of the sample vial and matrix modification to release the analytes from the sample matrix. This approach requires a further step to be optimised. Namely, the selection of the optimal trap adsorbent(s) that both traps and releases the analytes quantitatively with no break-through, irreversible adsorption or catalytic breakdown, while providing the best recovery of the analytes. It is also advantageous to choose a selective adsorbent that does not trap the matrix, for example a hydrophobic adsorbent for aqueous samples (more detail on sorbent selection is given under thermal desorption in this chapter). The optimal adsorption and desorption temperatures and the flows must also be determined and care must be taken to minimise activity, dead volumes and cold spots through the more complicated flow path and valve.

Purge-and-trap (P&T) analysis is even more sensitive down to ppt levels. It is very similar to DHS, however rather than the gas flowing over the top of the sample to remove the HS, the inert gas bubbles through the liquid sample to actively remove analyte molecules and sweep them into the trap. Care must be taken to prevent foaming of the samples which would result in contamination of the system and the trap. In newer instrumentation foam sensors are used and an anti-foaming agent can be added to reduce the likelihood of occurrence.

It is important to differentiate between the need to analyse a solid sample (which is a mixture or a single substance) and the need to extract the volatile and semi-volatile analytes contained within a solid sample.

Some solid-phase samples must be sampled in situ, for example volatiles emitted from building materials, car interiors or plants. For others, a sub-sample can be taken back to the lab and a portion of this analysed, for example soil, drugs, washing powders, meteorites and food.

Solid samples can be prepared in a similar fashion to liquid samples. Care should be taken to select a solvent which dissolves all of the sample as rapidly as possible. As with liquid samples, the choice of a suitable solvent is very much dependent on the sample properties. A solvent should be chosen which dissolves the sample, is amenable to gas chromatography (see Section 1.2.2) and in which the analyte is stable for the time required for the analysis. It may be necessary to transform the analyte into a more volatile substance to make GC analysis possible. Further details on derivatisation (Section 1.4) and pyrolysis (Section 1.3.2) are provided later in the chapter.

As with liquids, storage prior to analysis under controlled conditions is important to ensure the sample presented for analysis is representative. This may require the solid sample to be protected from light, water ingress and extremes of temperature, which along with other factors may alter the solid prior to analysis. Particular care must be taken with biological samples.

Treatment of the solid to maximise solubility in the chosen solvent may include milling the sample to decrease particle size and to increase the surface area, or use of sonication to encourage rapid solubilisation.

Solid samples, for which there is a need to analyse the volatile analytes they contain but for which no suitable solvents exist, can be prepared using other treatments including: HS analysis (see Section 1.2.6), TD and pyrolysis.

Thermal desorption is described for the concentration and transfer of gas phase samples to the GC-MS for analysis in Section 1.1.3.1. Small solid or viscous liquid samples can be directly thermally desorbed, at temperatures up to 350 °C so that no chemical bonds are broken, by placing them in a conditioned TD tube. Larger solid samples can be placed in a micro-chamber or macro-chamber up to 1 m³, in which they are heated and an inert gas sweeps the analytes into a packed TD tube. A schematic of the process is shown in Figure 1.3.

The parameters that need to be optimised are:

• Sample amount (weight) and size: large differences in the surface areas for the same sample weight will affect the time taken to fully desorb the sample.

• Dry purge: to remove water (and air) as water will produce a large vapour volume leading to contamination of the system.

• Pre-purge: the TD tube must be purged with carrier gas to remove oxygen and prevent oxidation of the sample.

• Desorption temperature: hot enough to extract the analytes but not too hot that the matrix carbonises and absorbs the analytes or causes reactions.

• Desorption time: the full sample needs to be heated and all volatiles extracted for reproducibility.

• Cold trap: the sorbent material used (see Section 1.1.3.3), trapping and desorption temperatures and time.

• Split flows: from the TD tube to the cold trap, then from the cold trap to the GC column. For low concentrations, a small split from the cold trap to the GC column is always advisable to produce better peak shapes, rather than using a splitless transfer.

Pyrolysis can be used to gain information about solid samples through thermal decomposition of the sample in a controlled environment (usually inert). The sample is rapidly heated to temperatures above 350 °C, usually in the region of 600–1400 °C, in which chemical bonds are cleaved within the macromolecular structure producing low molecular weight, more volatile analytes that are specific units of that macromolecule. These are volatile and can be transferred to the GC-MS for analysis, either through a heated transfer line or directly from a pyrolyser on top of the GC inlet. The collection of analytes can give an insight into the sample. Pyrolysis chromatograms, called pyrograms, are typically complex with many individual analytes obtained, the result produced is dependent on the pyrolysis temperature, heating rate, gas type and flow used for the inert atmosphere and the sample. The pyrogram obtained is a fingerprint of the sample and the profile or identification of the key analytes can be invaluable for particular applications such as the forensic analysis of paints or pigments or other high molecular weight analytes for which direct analysis is difficult, such as plastics, synthetic polymers, rubber, rocks, oil or wood. The pyrolysis products of polymers can give an insight into the composition of the polymer. In some cases, the polymer may simply depolymerise to yield the monomer or monomers. In other cases, this process may yield more complex products.

The mechanism of rapid heating varies depending on the type of pyrolyser. Common types include: inductive heating, resistive heating, a furnace, laser and heating within non-specialist pyrolysers such as within a programmable temperature vaporiser (PTV) inlet.

All pyrolysers share a common aim, to rapidly heat (20–1500 °C ms⁻¹) to a high temperature 600–1400 °C, with a short temperature rise time (TRt) and to hold for a short pyrolysis time (2–60 s), sufficient to fully pyrolyse the sample, but not produce secondary reactions. An example of a temperature-time profile (TTP) is shown in Figure 1.4a. Pyrolysis methods include:

• Isothermal: analysis at a single temperature with a single pyrogram produced. Includes flash pyrolysis with a TRt of approximately 8 ms.

• Sequential: the sample is heated to the same temperature multiple times with a pyrogram produced for each pyrolysis. This method is useful for kinetic studies to understand degradation and formation rates and activation energies. It is also used to determine if different substances give the same pyrolysis products at different rates.

• Fractionated: the sample is heated to different increasing temperatures with a pyrogram produced at each temperature. Used in complex samples and for thermally labile analytes. An example of fractionated pyrolysis of toner at three different temperatures is shown in Figure 1.4b, along with their TTP in 1.4a. 

• Pyrotomy: the sample is pyrolysed for a very short time (20 ms) to only pyrolyse the surface of the sample in contact with the filament. Used for non-homogenous samples in which the sample can be analysed layer by layer.

Sample preparation is very straight forward, microgram to milligram quantities (pyrolyser dependent) of a sample are placed onto a ribbon filament or into a quartz tube or metal cup which is then sealed into the pyrolyser and rapidly heated or dropped into a furnace. Sub-pyrolysis temperatures may also be employed to thermally desorb any volatile species which may be present within the matrix, such as plasticisers or other additives. Pyrolysis parameters that need to be considered are:

• Sample size: a small sample size, which has a constant weight and shape, is essential to ensure that all of it degrades fully and rapidly otherwise it may result in poor reproducibility. Samples are rarely homogenous, therefore quantitation can be difficult.

• Pyrolysis temperature: depends on the sample and problem to be solved. Higher temperatures can produce larger, analytically more significant molecules. Too high and the sample degrades producing uncharacteristic information with simultaneous bond breakages to create very small and non-specific free radicals and molecules, carbonisation can also occur. Too low, and the sample may not be totally pyrolysed. Different samples have different degradation rates and can produce exothermic or endothermic reactions resulting in the pyrolysis temperature being different to that set in the unit. To optimise the pyrolysis temperature, the analyte area should be plotted against the pyrolysis temperature.

• Pyrolysis time and ramp rate: long pyrolysis times, including slow pyrolysis ramp rates, can produce secondary reactions in some samples, resulting in analytes being identified that were not originally part of the sample structure.

Pyrolysers can also be used for derivatisation, for example thermochemolysis (thermally assisted hydrolysis and methylation) is used to selectively break ester and ether bonds, before the products are methylated and analysed. The sample can then be further analysed by pyrolysis.

The main disadvantage of this technique is that often the chromatograms are very complex and contain many analytes to identify, making it challenging to reconcile the information obtained with the sample. However, if time is taken to construct reference libraries and deconvolute the complex GC-MS dataset this technique can provide valuable information.

There are a wide number of liquid extraction techniques for the preparation of solid-phase samples, the more common ones include reflux, soxhlet and static extraction, either in an open or sealed vessel.

Each of these techniques relies on the action of a solvent on a solid sample, the solid matrix itself is insoluble in the selected solvent but the constituent volatile and semi-volatile analytes are soluble and can be extracted by the action of the solvent penetrating the solid matrix over time.

This is perhaps the simplest form of liquid extraction. Here, the solid sample is simply placed into the liquid solvent and left for a fixed period of time to allow the solvent to penetrate the solid and extract the volatile and semi-volatile constituents through diffusion. To aid this process, and indeed any of the described extractions from a solid sample, it is possible to increase extraction efficiency by:

1. Increasing the surface area of the solid undergoing extraction. This is typically done by grinding or cutting the solid into smaller particles or parts. If volatiles may be lost, cryogenic grinding (also known as freezer grinding or milling) can be employed using liquid carbon dioxide or liquid nitrogen.

2. Increasing the temperature of extraction, this increases the diffusion rate of the volatile and semi-volatile analytes from the solid into the solvent.

3. Increasing the time for which the solid and solvent are in contact with each other. Diffusion can be a slow process and will become slower as the equilibrium is reached between the concentration of the analyte in the solid and the solvent. It may be necessary to repeat the extraction of the solid with fresh solvent to remove all of the analytes.

Variations of this also include the use of sonication (sound energy) from acoustic cavitation owing to the formation, growth and implosive collapse of bubbles in a liquid, this can be used to speed up the extraction process in several ways. The samples are normally placed in a laboratory ultrasonic bath (filled with water), the extraction solvent is placed in a beaker or a round bottom flask which is secured within the ultrasonic bath. A high frequency transducer in the kHz range causes cavitation of the water in the bath and the extraction solvent. This energy speeds up the extraction both through agitation and heating of the extraction solvent. The disadvantage of this method is that the energy levels can be hard to control and therefore the robustness and reproducibility can be poor if the energy levels vary.

Another variant is the use of microwave energy to heat the extraction solvent in a sealed container above its boiling point. The effect of elevated pressures and temperatures can speed up the extraction of analytes. Different solvents will be heated by the microwave energy more or less, depending on their dipole moment. The solvent dielectric constant can be used to select extraction solvents that interact with the microwave energy the most and thus are quickly heated. Generally, the more polar the solvent the more efficiently it heats, thus water is heated very quickly and hexane not so quickly.

Microwave extraction can be influenced by a number of factors which are common to the other extraction methods described here such as the extraction time, temperature, microwave power and the solid sample characteristics, however, the solvent choice is the most important. Solvent mixtures can be used to optimise the characteristics of the extraction including the addition of salts and hexane/water or acetone/hexane mixtures to improve the extraction efficiency and ensure complete extraction. The hexane swells the solid matrix and solubilises the non-polar analytes, the acetone interacts with the microwave energy to heat the extract. Clearly some quite complex optimisations are required once mixtures are introduced, but this makes microwave extraction very flexible. Microwave extraction units also allow precise monitoring of microwave energy levels, temperatures and pressures within the sealed unit containing the extraction solvent and sample, leading to much greater control over the extraction conditions.

In reflux extraction, the solid sample is placed into a round-bottom flask, the solvent is added and a refluxing condenser is placed on top. The flask is heated to above the boiling point of the solvent, the solvent boils and is returned to the round bottom flask via the condenser. Over time this refluxing action removes and collects the volatile and semi-volatile analytes. After a fixed period of time the solvent is allowed to return to room temperature and is then available to be analysed as a liquid extract using GC-MS. Extraction times can be quite long, with periods of 16 hours or more not being uncommon, depending on the nature of the solid, the analytes and how complete and reproducible an extraction is required. It is difficult to process large number of samples, with long extraction times plus the glassware, heat source and condenser set required for each sample limits the number of samples that can be processed in parallel. The volume of solvent used is also quite high, with associated safety aspects, costs and waste disposal to consider. The supply of water may also be an issue and all these factors contribute to the environmental impact of analysis.

Soxhlet extraction is a similar process to reflux extraction with the difference being that the solid sample is suspended above the solvent in a round bottom flask, held within a specialised type of condenser, the soxhlet condenser. Therefore, the solid sample is only exposed to the condensing solvent after it has evaporated from the round bottom flask, rather than the boiling solvent. This means that the temperature of extraction is slightly lower and the solid is always exposed to just boiled pure solvent without the extracted material, which collects in the flask. The technique potentially improves the extraction efficiency as the extraction media is never saturated by the extractants. Soxhlet extraction shares many of the same characteristics of reflux extraction, it is simple and inexpensive to setup and is an effective extraction technique if time is given for extraction to occur. Its disadvantages are the speed of extraction and a limited ability to process large volumes of samples in a timely manner. It also uses relatively large volumes of solvent, which adds complexity if the concentration of the analytes in the solid are low and are therefore further diluted by the extraction procedure. Therefore, solvent concentration may be required, see Section 1.5.

Pressurised fluid extraction (PFE) is a generic term for a technique which uses high temperatures and pressure to reduce the extraction time. It includes subcritical and supercritical fluid extraction.

Accelerated solvent extraction (ASE) is a form of PFE and is a registered trademark of Dionex. PFE extraction systems use temperatures of up to 200 °C and pressures of around 1500 psi within an extraction cell. The extraction cell contains the solid sample and the solvents are introduced either using static extraction or a dynamic extraction step and often a combination of the two. The temperature and pressure reduce the extraction time and reduce the volume of solvent needed to extract the solid. As with microwave extraction the extraction is more controlled owing to precise monitoring of the temperature and pressure, but unlike microwave extraction, it is possible to dynamically adjust the extraction solvent to optimise the extraction process. There are numerous parameters which affect the extraction making it necessary to undertake careful optimisation to achieve a reliable and effective result, when properly optimised it is a very effective technique!

Supercritical fluid extraction (SFE) is a sub-category of ASE. This extraction technique uses the physical properties of supercritical fluids to improve the extraction. A supercritical fluid is a substance above its critical point. A fluid above its critical point has properties which enhance the extraction of solids such as enhanced effusion into a solid (gas-like) and the ability to dissolve solids (liquid-like). A substance is brought above its critical point when pressure and temperature are applied. A common fluid is supercritical carbon dioxide because its critical point is within a convenient pressure (74 Bar), and temperature range (31 °C) and it is easily available (Figure 1.5). Many other fluids can be used but they have practical issues relating to the toxicity and availability. Carbon dioxide is frequently modified during the extraction process by the addition of co-solvents such as methanol. In SFE, the fluid is pumped through an extraction cell containing the solid to be extracted. As with microwave extraction and ASE the evaluated pressure in a sealed container enhances the extraction reducing the extraction time.

As you might expect each extraction technique has its strengths and weaknesses these are summarised in Table 1.1. Which technique is selected will depend on the application as well as these factors.

Derivatisation is used for various reasons in GC-MS analysis and these can be sub-divided into four main groups:

1. To make the analytes more volatile, and thus amenable to gas chromatography.

2. To remove functional groups which make gas chromatography problematic, for example hydroxyl, carboxyl or amine groups.

3. To improve detectability, this may overlap with point 2 to a degree if functional groups are changed to improve the chromatography. Furthermore, alternative functional groups may be formed which generate a better response in the detector.

4. To improve the thermal stability of the analyte to facilitate GC-MS analysis.

The general approach is to chemically transform the target analyte(s) through the use of a reaction with a derivatisation reagent either prior to the subsequent analysis as a separate sample preparation step or in the injection port of the gas chromatograph. PTV injectors are useful for this. There are four commonly employed derivatisation modifications: these are silylation, acylation, alkylation and esterification.

There are many different reagents which have been developed and these have been reviewed and discussed elsewhere.

Silylation aims to replace an active hydrogen on an OH, SH or NH group in a molecule (Scheme 1.1). This has the effect of increasing the volatility and reducing the interaction with the GC flow path residual silanol groups, thereby reducing peak tailing.

The ease of silylation follows predictable trends as per nucleophillic substitution and the basic nature of the leaving group on the silylating agent (Table 1.2)

Acylation replaces an active hydrogen with an acyl group (Scheme 1.2). This increases volatility and can also improve thermal stability and reduce interaction with the active sites through the GC (Table 1.3)

Esterification is a subset of alkylation in which organic acids are reacted with alcohols to produce the corresponding ester. A common example of this is fatty acid analysis owing to their corresponding methyl esters (FAME analysis). Transesterification is used to displace one ester group with another to produce a new ester. This is commonly deployed in the analysis of fats and triglycerides (Table 1.4).

Sample preparation has the advantages of removing the matrix or concentrating the analytes, however this process can also cause the loss of analyte molecules. Analyte recovery is an important parameter to check when developing and validating a method. Sample preparation is the largest source of error in the analysis of a sample by GC-MS, therefore it is important to optimise the entire GC-MS method, including sample collection and sample preparation. The recovery criteria often dictates that it should be more than 80%, however if the accuracy and precision of the method are good, then a lower percentage recovery may be acceptable. Reproducible recovery indicates that the method is robust.

Recovery can be calculated by:

In which Area A is equal to the peak area of the prepared sample (which could be an extraction) and Area B is the peak area of the prepared (extracted) sample matrix in which the analyte was added after the sample preparation step (post extraction).